Actual benefit of Kitsune
The paragraph you quoted from W20 Changing Breeds says that:
- Kitsune can "learn potent Gifts of their own", without a teacher.
- Kitsune can "learn the Gifts of other Changing Breeds" (any of them), provided that they find a teacher.
The experience cost is specified as "the same as for a non-breed/auspice/tribal Gift", and we don't know if it applies to both points and Kitsune have to pay increased cost if they learn a "potent Gift" on their own.
I do not know if “potent” is just a fluff word here or a game term; as it is not capitalized, I would assume the former.
Let us compare it to the “standard situation of W20” — to the Garou.
Garou, according to the Learning Gifts paragraph on pages 151-152 of W20 Core, always need a mentor to learn a Gift. If it is a Spirit, the process goes fast, if it is another Garou, not so fast, but still possible. Breed, Auspice and Tribe Gifts cost 3*level XP, others cost 5*level XP.
Garou do not get the opportunity to learn the Gifts of other Changing Breeds (other Fera). At least I was not able to find something that allows them to do it, though neither was something directly forbidding it found. The sentence "a werewolf may learn the Gifts of other breeds, auspices or tribes" implies not being able to learn Kitsune Gifts, but rather Gifts of Homid if you are Lupus, for example.
So, in W20 the benefit of Kitsune is not “quantity of Gifts” (the ability to learn more Gifts from their list with the same amount of XP), but rather “quality of Gifts” (the ability to learn more types of Gifts and easy access to Kitsune Gifts). This is further supported by the phrase "Kitsune rely heavily upon their magical strength and versatility."
White Wolf books are flawed
The books printed by White Wolf inherit a lot of text from each other, and when rules (or lore) change from edition to edition, some now unrelated or obsolete text may occasionally stay.
For example, V20 Core, p. 107:
If you don’t have any dots in a Knowledge, you cannot
even attempt a roll involving it unless the Storyteller
gives explicit permission (such as where common trivia
is concerned). If you don’t know Spanish, you can’t try
holding a conversation en español on your Wits alone.
But Linguistics is no longer a Knowledge in V20, languages are 1-dot Merits, and you either know it or don't know. This probably still explains how do Knowledge checks work, but the example check is irrelevant. If they changed it to, for example, a Computers check, it would be a lot more relevant, and in addition it would explain what does it mean to have 0 dots in Computers: not being able to use computers at all, or only to hack into the systems?
Or, regarding W20, see the question "How does Calm Heart work?": a Merit from VtM series was directly reprinted without fixing the issue with vampire and werewolf Frenzy functioning differently. You would not be able to use this Merit out of the box, you would have to home-rule some other way to make resisting Frenzy easier or find another book to use.
And the most iconic example being Page XX.
This makes understanding WW books... hard. Sometimes. At least do not expect it to always be easy.
GM fiat is what really limits you ever
You asked about the GM fiat -- here is my answer.
Having to find a teacher to learn a "foreign" Gift effectively means that GM can deprive you of any Gift. He/she may also deprive you of any other option you want to take, just because "Rule 0" gets reprinted in WW books over and over and remains essential, allowing to rule however your ST wants to rule. This is especially important regarding the ability not to look for a teacher when learning a native Kitsune Gift, my inner oWoD player senses tell not to expect every ST to actually allow learning any Native Gift that way.
Whatever is written in the book, all of the GMs that I met had something work very differently compared to RAW, sometimes intentionally (they did not like how it is done by WW), sometimes just because they didn't bother to learn all of the rules. And I also noticed that the more people call some Storyteller I know good, the more are the rules of that Storyteller different from RAW. That is a thing to consider.
Just because it is (un)fortunately not Pathfinder, even if it is possible to do something in oWoD according to RAW, one should be totally ready hear a "no, just no" from the Storyteller.
Note: The calculation in normal text is based on a lunation. The corrected numbers for a sidereal month are found in the supertext. Note that approach has an in-built self-correction: When pinning the full moons to the known dates of a lunar calendar, there is a scaling process included. This compensates the 51-hour error between synodic and sidereal month.
Pretext
This is mainly a question of mathematics. A moon in 27.3/29.5 days long. Divided upon 5 Auspices, this is 5.46/5,9 days per Auspice. But how to distribute these? We can't place them all one after another in the listed order, as then the new moon follows the full moon, and those should be half a month apart! We need to distribute a little...
Graphical Method
Generating the 'bar'
Let's start in the middle of a full moon phase, as we can't cut that apart but for marking the center - the absolute full - of it. And now we draw out 2.73/2.95 days, as that is half the phase's length...
Now, we repeat these for the Gibbous (allmost full), Half and crescent Moon, stacking them...
Next comes the new moon, which again can't be divided into a first or second half of the month, so it will be a full 5.46/5.9 days. Then again the phases with 2.73/2.95 day in backward order.
Calender dating
Note the gap at the end to the 28 day/4 week display - this is responsible for the shifting of the phases. But taking this bar and a moon calendar, it is easy to place each day towards one of the phases, and with more math, one could decide exactly when in the night/day the auspices shift.
Example: Tuesday 3rd July 2012 was a full moon, the 18th a new moon and 2nd of August a full moon again according to my moon calendar. Using a properly scaled bar like the one before (with a full, 5.465.9 days full moon at the end) the lunar month looks like this:
Full Math Approach
Taking a Sidereal Month as a base we know that an Auspice is supposed to be 5.9 days or 141.6 hours or 141 hours and 36 minutes.
Let's get the full moon moment of the 3rd July 2012: according to lunaf.com that is 18:52, as the center of the Time. WA says this time has 99.92%, so close enough.
2 days, 22 Hours, 48 Minutes later we should change from full moon to Gibbous Moon. That is 6th of July 2012, 17:40 UTC. 2.95 days till each new phase starts - of in the case of the New and Full moon, reaches its zenith. That can be easily taken to excel or another spreadsheet program! Fill in one date for a full moon in B2
and in the field below add =B2+2.95
the 2.95 is assuming average moon length - an error! - then pull it down to fill the column. The categories are to be in the following order: Full Zenith, Gibbous Start, Half Start, Crescent Start, New Start, New Zenith, Crescent Start, Half Start, Gibbous Start, Full Start, Full Zenith. The last entry is again the first entry of the next lunar month, so we get a repetitive pattern, and don't even need to scale anything1.
Compensation for variable moon lengths
Our calculation gives 2nd of August 6:52 UTC as full moon zenith, lunaf.com gives 03:27 UTC. The error of 3 hours comes from the variance the synodic month has, ranging from 29.18 to 29.93 days, resulting in Auspiece-lengths of 2.918 to 2.993 days. The chosen month in the example is mathematically 29.358 days long.
Because of this it is advisable to look up the dates of full moons in the years played and then calculate the correct sidereal lengths to correct for this error. The formula gets considerably more complex though. So take a look at this spreadsheet that fixes the sidereal month's length by taking input in the shape of a start and end date for a month.
Illumination + lunar calender method
Another variant would be to look up the illumination of a moon for certain days. The Full moon of course has 100% illumination, the new moon 0%. The swing from full moon to new moon and back to full moon thus swings twice over the scale, a total of 200 percentage points. Each phase thus should cover 200/5=40 percentage points, 20 of these on the waning and 20 on the waxing side. These illuminations can be looked up at some online tools, and then easily compared to a simple chart that does not care for which half in the lunar month one is in, as it mirrors around the 0 and 100 point to swing back:
- 80-100% - Full
- 60-80% - Gibbous
- 40-60% - Half
- 20-40% - Crescent
- 0-20% - New
Note though that illumination does not exactly follow a perfect distribution over the month and visible illuminated area varies depending on the position on the globe. Check where your calendar is dated for, or use Wolfram Alpha to a local position!
Comparison & Verification
Let's run a test: looking up the 27th July of 2012
The Illumination+Calender method shows at lunaf.com or Wolfram-Alpha that at noon it was a 62.01% illumination, which is a Gibbous moon in the waxing half of the lunar cycle.
Comparing this to the other methods we see that the moon just changed into this phase on the 27th (graphical Method) and did so at 9:16with long-term median length / 6:32corrected for month length AM. Close enough, aye? All three methods came to a workable result for the example day!
Careful checking on WA shows, that the actual 60% point was at 7:43 AM in Greenwich at 51°28′40″N 0°00′05″W. This error is relative minuscule and is to a good degree derived from different full moon times.
Best Answer
The first thing to do is dispense with the idea of "dodging outside of your Initiative order." The only time you'd need to dodge is outside of your Initiative order, since if you're dodging, you're not the acting party. What's most often being referred to in those circumstances is "dodging when you don't have an action dedicated to defense." In those cases, you can make a Willpower roll (or spend a Willpower) to change a previously declared action to a "dodge." (The combat system, remember, presumes that you'll assign actions in Initiative order and resolve them in reverse order. In a forum, you may not be using that rule.)
In Werewolf: the Apocalypse 20th Anniversary, the use of Willpower to convert a previously declared action into a defense is on page 289.
In Werewolf: the Apocalypse Revised, aborting to a dodge has no cost, as per page 205 and 206. "A character can choose to take one of three defensive actions as well. She may also choose to abort a previously declared action in order to do so.[…] Normally, once a player has declared an action, she may not change it. If she has a good reason to do so (a packmate kills her character’s intended target, for example) she may change her action, but she adds one to the difficulty. Aborting to a defensive action does not change the difficulty of said action.")
Now, if you have no more actions left in your turn to spend, you have nothing to abort or change and may not dodge. This is a very good reason to spend Rage on extra actions at the start of your turn.